marcan changed the topic of #asahi to: Asahi Linux: porting Linux to Apple Silicon macs | "Does XXX work yet?": https://alx.sh/fs | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Topics: #asahi-dev #asahi-re #asahi-gpu #asahi-alt #asahi-stream #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
<chadmed>
eris: many people have been dailying it for a long time
<eris>
alright thank you so much i could not find info on this :)
<chadmed>
https://alx.sh/fs this page will tell you what works and what doesnt. for all intents and purposes, at this stage using these machines daily with linux is just like using any other linux machine
espo has quit [Ping timeout: 480 seconds]
espo has joined #asahi
as400 has quit [Remote host closed the connection]
as400 has joined #asahi
<eris>
oh ok
<eris>
it doesnt have enough for me yet but i cant wait to try it!
ptudor_ has joined #asahi
ptudor has quit [Ping timeout: 480 seconds]
gabuscus has quit []
Brainium has quit [Quit: Konversation terminated!]
jeisom has quit [Ping timeout: 480 seconds]
<chadmed>
eris: what's missing?
gabuscus has joined #asahi
atipls has joined #asahi
marvin24_ has joined #asahi
marvin24 has quit [Ping timeout: 480 seconds]
cylm has quit [Ping timeout: 480 seconds]
c10l has quit [Ping timeout: 480 seconds]
c10l has joined #asahi
mps has quit [Ping timeout: 480 seconds]
brunsgaard has quit [Ping timeout: 480 seconds]
SalimTerryLi has quit [Remote host closed the connection]
mps has joined #asahi
ptudor has joined #asahi
ptudor_ has quit [Ping timeout: 480 seconds]
tano70_Air has joined #asahi
crabbedhaloablut has joined #asahi
<tano70_Air>
Hello, Asahi install work with new Mac Sonoma ?
<eiln>
tano70_Air: yes, you can install sonoma
<tano70_Air>
Ok! Thanks eiln
brunsgaard has joined #asahi
brunsgaard has quit [Ping timeout: 480 seconds]
jeffmiw has joined #asahi
brunsgaard has joined #asahi
ciggi has quit [Read error: Connection reset by peer]
ciggi has joined #asahi
ciggi has quit [Read error: Connection reset by peer]
ciggi has joined #asahi
brunsgaard has quit [Ping timeout: 480 seconds]
brunsgaard has joined #asahi
brunsgaard has quit [Ping timeout: 480 seconds]
lewurm has quit [Ping timeout: 480 seconds]
mohit815 has quit [Quit: mohit815]
mohit815 has joined #asahi
flom84 has quit [Remote host closed the connection]
mohit815 has quit []
mohit815 has joined #asahi
lewurm has joined #asahi
flom84 has joined #asahi
lewurm has quit [Remote host closed the connection]
lewurm has joined #asahi
flom84 has quit []
flom84 has joined #asahi
brunsgaard has joined #asahi
brunsgaard has quit [Ping timeout: 480 seconds]
flom84 has quit [Quit: Leaving]
brunsgaard has joined #asahi
brunsgaard has quit [Ping timeout: 480 seconds]
<j`ey>
is everyone seeing: tainted: local-hwclock, in systemctl status, or do I have something misconfigured?
<janneg>
j`ey: not seeing that. there is however something odd about the rtc, it registers later than expected. in my current boot it registers only after dcp was probed although it is built-in and has as far as I can tell no dependency on any module
<j`ey>
hm ok, thanks. I also get messages from systemd that the clock jumped back, I guess that's due to that problem then
SalimTerryLi has joined #asahi
SalimTerryLi has quit [Remote host closed the connection]
SalimTerryLi has joined #asahi
alarumbe has quit [Ping timeout: 480 seconds]
brunsgaard has joined #asahi
dhmltb^ has quit [Remote host closed the connection]
atipls has quit [Quit: ZNC 1.8.2+deb2build5 - https://znc.in]
melonai339 is now known as melonai
roxfan has joined #asahi
Fischmiep has quit [Remote host closed the connection]
Fischmiep has joined #asahi
CoolStar has joined #asahi
cylm has joined #asahi
rvalue has quit [Remote host closed the connection]
rvalue has joined #asahi
coderobe has quit [Quit: Ping timeout (120 seconds)]
linuxgemini has quit [Ping timeout: 480 seconds]
linuxgemini has joined #asahi
Moprius has joined #asahi
coderobe has joined #asahi
flom84 has joined #asahi
kidplayer_666_ has joined #asahi
<kidplayer_666_>
eris: Hey there! I just started using fedora with Gnome recently, and the state it's already in is very impressive
<kidplayer_666_>
Im dayliing a MacBook pro m1 with a touchbar
<kidplayer_666_>
And it's awesome
<kidplayer_666_>
However
<kidplayer_666_>
Several issues remain
<kidplayer_666_>
The touchbar doesn't correctly go to sleep and the battery is quite a bit worse, meaning it ain't quite there yet for me to daily drive it
<kidplayer_666_>
And speakers are still not there
<kidplayer_666_>
But the progress is still staggering
<kidplayer_666_>
And im constantly coming back to linux to tinker
<kidplayer_666_>
And check the progress
<kidplayer_666_>
Much of the software I need already works
flom84 has quit [Quit: Leaving]
<kidplayer_666_>
Although im not the most intensive user
<kidplayer_666_>
But stuff like thunderbird (through betterbird), VSCode, discord (through Armcord) all seem to work decently. Even WhatsApp has a usable client!
kidplayer_666_ has quit [Remote host closed the connection]
kidplayer_666_ has joined #asahi
flom84 has joined #asahi
kidplayer_666_ has quit [Remote host closed the connection]
Moprius has quit [Quit: bye]
chadmed has joined #asahi
dextaa6 has joined #asahi
<j`ey>
can someone with fedora `cat /etc/adjtime` and say what the 3rd line is? if it exiss
<j`ey>
exists
<cylm>
@j`ey It says for me LOCAL
<cylm>
but not on asahi currently, just random fedora install. is that fine?
ydalton has joined #asahi
<ydalton>
j`ey: for me it's LOCAL
<j`ey>
thanks
<j`ey>
ydalton: can you run systemctl status, does it say 'tainted: local-hwclock'?
<ydalton>
eris: i've been dailying it for 2 weeks now for school, and it's alright,
<ydalton>
but i am missing some features like proper x86 emulation, virtualization, but
<ydalton>
nothing deal breaking really
<j`ey>
so the fedora qemu package didnt work for you either?
<ydalton>
j`ey: yep, right above CGroup right?
<j`ey>
ydalton: yeah
<ydalton>
yeah, it does indeed say "Tainted: local-hwclock"
<ydalton>
j`ey: i don't know, kvm keeps shitting itself for some reason
mohit815 has quit [Quit: mohit815]
<ydalton>
with systemd-nspawn, so that's something
<ydalton>
i needed vagrant for school but i was able to replicated the same effect
mohit815 has joined #asahi
<ydalton>
oh my god, i have to disable auto-fill-mode in emacs for this buffer
dextaa6 has quit [Quit: Leaving]
ydalton has quit [Remote host closed the connection]
dextaa6 has joined #asahi
ydalton has joined #asahi
<ydalton>
okay, i changed my emacs config and i hope it's fixed
<ydalton>
j`ey: i read about a workaround by adding "highmem=off" in the qemu command line, but it did not work, so i'm out of ideas
<ydalton>
did you try virtualization at all?
alarumbe has joined #asahi
tano70_Air has left #asahi [Leaving]
ydalton has quit [Quit: ERC 5.5.0.29.1 (IRC client for GNU Emacs 29.0.92)]
<j`ey>
ydalton: that workaround was needed for older qemus
<j`ey>
ydalton: you need to provide errors / commands you ran
ydalton has joined #asahi
tano70_Air has joined #asahi
eiln has quit [Quit: WeeChat 4.0.4]
<ydalton>
the fuck, virt-manager works?
<ydalton>
and it says domain kvm so it's using kvm
<j`ey>
janneg: I wonder if the devlink errors around smc/rtc have anything to do with it. I tried with fw_devlink=off, but then I just get a blank screen, so that wasn't a good start! i don't have a kernel build setup, so can't do a lot to investigate currently
<j`ey>
display goes blank even with nomodeset
brunsgaard has quit [Ping timeout: 480 seconds]
<janneg>
"fw_devlink=off pd_ignore_unused" and disable runtime pm
mspe has joined #asahi
<j`ey>
is there a commandline option to disable runtime pm? can't see one in the Docs, pd_ignore_unused got the screen to stay on, but it seems to have just drawn a single image and thats it
<j`ey>
it can probably wait until I get a kernel building agian, just thought it would be fun to dig a bit
tano70_Air has left #asahi [Leaving]
<Tramtrist>
following marcan on mastodon almost makes the platform on its own...
i509vcb has quit [Quit: Connection closed for inactivity]
brunsgaard has joined #asahi
tano70_Air has joined #asahi
mspe has quit [Quit: WeeChat 3.0]
mspe has joined #asahi
novafacing9926 has joined #asahi
brunsgaard has quit [Ping timeout: 480 seconds]
ydalton has joined #asahi
<ydalton>
man marcan is something else with his rants
<ydalton>
but he is right, apple's teams are pretty incompetent
<ydalton>
from the speakers damage thing to the gpu firmware exploit, h264 decode in the macOS kernel driver, and now this shenanigans
<ydalton>
i just love this
<ydalton>
if i understand this correctly, every 1 in 128 samples it goes to zero, creating this weird clipping sound
ydalton has quit [Quit: ERC 5.5.0.29.1 (IRC client for GNU Emacs 29.0.92)]
ydalton has joined #asahi
<ydalton>
or not zero it out, just not process that 1 in 128 sample
ydalton has quit []
<eris>
chadmed: the stuff that is missing is really the speakers but i can use my headphones for that
brunsgaard has joined #asahi
kaazoo has left #asahi [#asahi]
kaazoo has joined #asahi
chadmed has quit [Ping timeout: 480 seconds]
<Tramtrist>
m2 air speakers are also what Im waiting for :)
brunsgaard has quit [Ping timeout: 480 seconds]
brunsgaard has joined #asahi
AnuthaDev has joined #asahi
<sven>
if the teams are so incompetent how did they manage to create such an awesome machine?
tano70_Air has left #asahi [Leaving]
<sven>
it’s much more likely they are under immense time pressure and/or have different priorities
ydalton has joined #asahi
<ydalton>
well, in that case, then it's the management that are pretty incompetent (like marcan said)
<sven>
apple prints money, how is that incompetent management then?
<ydalton>
and if so, i feel sorry for them
<sven>
bugs just happen, and many of them are just dumb
chadmed has joined #asahi
<sven>
I’m sure if you go through m1n1 or Linux it whatever you’ll find equally dumb things
<ydalton>
sven: not incompetent enough to lose money i guess
<ydalton>
it is human nature to make dumb bugs, i know that first hand
<j`ey>
sven: no way m1n1 is perfect!
<ydalton>
i'd love to believe the rumor apple employees use m1n1 over their tooling because its better
<sven>
sure, because we invested lots of hours into making m1n1 that good
<sven>
trying to get funding for that many man hours for a project that no immediately obvious ROI is much harder than people deciding they want to build it in their free time
<ydalton>
true i guess
<sven>
as I said, if I went through the m1n1 or Linux commit history I’m pretty sure I could also easily find something that makes us look incredibly incompetent
<ydalton>
well, you probably remember the time the X.org hack in the drm subsystem went viral
<ChaosPrincess>
ydalton: its just how big corps work at the end, even if you have an ancient ticket with a "we should fix this code", even if it has exact instructions on how to do it and why it is bad, it still ends up deprioritized in favour of features some manager dreamed up a day ago.
<sven>
yeah, and that thing is isn’t dumb
<sven>
the xorg thing looks dumb without context
<sven>
but iirc it was either that or break I don’t know how many machines
<sven>
yet it’s easy to call whoever commited that incompetent without context
<sven>
I just don’t like people going around and calling other’s work incompetent because that makes for a pretty nasty environment imho
<ydalton>
i'm not sure we're talking about the same xorg drm bug but anyway
<sven>
In talking about the “if process name starts with “X” hack disable some feature” that made its rounds a few months ago
<sven>
*I’m
<sven>
was there another one? :D
<ydalton>
yeah that's the one
<ydalton>
and it was reverted afterwards
<j`ey>
we have that same hack :D
<sven>
yeah, and it looks dumb but it’s there for a good reason unfortunately:D
<ydalton>
well, if i read the code correctly, it's more like "refusing to let Xorg open the touchbar dri node"
<ydalton>
but anyway, nobody wants to touch Xorg anymore
<sven>
there’s a similar hack upstream that disables some(or maybe all?) atomic drm features for processes starting with X
<ydalton>
wait
<ydalton>
that's the one i was talking about
<ydalton>
"take the atomic toys away from X"
<ydalton>
is a very lovely title
<sven>
yeah
<ydalton>
wait, what was the one you were referring to then? this one?
<sven>
I’m confused now :D
<ydalton>
me too :P
<ydalton>
how many hacks need to check if the process's name begins with X to avoid dealing with Xorg???
<ydalton>
it's just amusing to me lol
<sven>
two I think, the upstream one and the thing j’ey mentioned that only exists in the Asahi trees
<ydalton>
in adp_open iirc
<sven>
j`ey I mean. Autocorrect doesn’t want me to hilight him ;)
<j`ey>
sven: I was young and dumb when I thought ` would be a good replacement for o :D
<sven>
:D
<ydalton>
oh, now i see :D
<j`ey>
sven: btw did you end up writing slides for your talk? maybe saving them for another conf or?
<sven>
oh, yeah, I have the slides almost ready
<sven>
would just have to polish them a bit more
<ydalton>
when's the day?
<sven>
still bummed I got sick just before the conference:(
<j`ey>
ydalton: it was last week, but sven was sick and couldnt make it
<ydalton>
damnnnnn
<sven>
yeah :/
<ydalton>
i was looking forward to it but hey what can you do
<ydalton>
hm, it was in paris i see
<sven>
yeah, not even that far from me.
<ydalton>
neither from me too
<sven>
I think it would’ve been a 4 hour train ride or so
<j`ey>
sven: oh, I thought you actually got the train there
<ydalton>
you live in france?
<sven>
I got sick on the day of the train ride :(
<sven>
Germany
<sven>
I guess that’s still better than getting sick once I had already been in the hotel though
<ydalton>
really? cuz i know a two hour train ride can get you from the coast of belgium to geel (city in belgium)
<ydalton>
they must be high speed trains then
<sven>
I live in southern Germany and it’s like an hour to the border from here
<sven>
it’s either ICE or TGV and both are pretty fast
<MichaelLong>
eris, yes speakers and video out for laptop, but I don't have hope about the latter
<ydalton>
i've never taken nmbs but i know people who have, and they've had a pretty shit experience with them
<ydalton>
actually i have taken them, but only a few times in my life
<j`ey>
MichaelLong: it will take time, but its coming!
<ydalton>
idk if we have high speed rain in belgium
<ydalton>
s/rain/rail
<ydalton>
there's definitely high speed rain in belgium XD
<ydalton>
j`ey: video decode he means right?
<j`ey>
'video out', external display that sounds like
<ydalton>
well, jannau got display out working on the m2 mini
<ydalton>
so that gives me hope
<ydalton>
external monitors aren't a deal breaker but clamshell mode would be nice :)
cylm has quit [Ping timeout: 480 seconds]
ydalton has quit [Quit: ERC 5.5.0.29.1 (IRC client for GNU Emacs 29.0.92)]
roxfan has quit [Remote host closed the connection]
crabbedhaloablut has quit []
i509vcb has joined #asahi
lynndotpy has quit [Quit: bye bye]
lynndotpy has joined #asahi
brunsgaard has quit [Ping timeout: 480 seconds]
jeisom has quit [Remote host closed the connection]
jeisom has joined #asahi
<leio>
I only have display out on my M2 Studio :)
<leio>
TIL clamshell mode
ydalton has joined #asahi
<ydalton>
leio: you've never heard of clamshell mode? well i guess it's a mac thing mostly
<ydalton>
and you're using llvmpipe right?
darkapex has quit [Remote host closed the connection]
darkapex has joined #asahi
<leio>
very soon, yes, finishing off final install now
<ydalton>
well that's nice
<ydalton>
expert install success stories i guess
<leio>
we don't talk about that here!
<leio>
I made my own installer to get gentoo going while alarm installer doesn't work on the new devices at all to go through the "livecd "method
<ydalton>
well it paid off
<stintel>
leio: rolled my own install medium too when I installed gentoo on my m2 air 6 weeks ago or so :)
<leio>
sorry, that sounded more than I did; I just rolled my own zip and such for the existing installer, the hard work is still ahead to brush things to usability to others
<stintel>
and boy am I happy with this thing. probably the best Linux on laptop experience I had in my life
<stintel>
despite not having audio and external display yet
<ydalton>
tbh even this macbook air is like the most powerful and portable linux laptop i've had
<ydalton>
i'll see how long i can daily drive this for school
<stintel>
this thing compiles some things faster than my 40c/80t workstation (ok it's a low clockspeed oldish CPU but still)
<stintel>
but I guess that means I need to look at upgrading my workstation :)
Enzime has quit [Remote host closed the connection]
<ydalton>
stintel: that's some beast
<ydalton>
i need to do proper benchmarks with my desktop vs this macbook
Enzime has joined #asahi
<stintel>
yeah with gentoo you can easily check your compile logs with timing :)
<ydalton>
i don't personally use gentoo but i am considering it for my desktop
<j`ey>
looks like chromium on m2 will be fixed on fedora soon
<ydalton>
stintel: but don't you have 40 cores on that thing tho
<stintel>
yes, dual Intel(R) Xeon(R) CPU E5-2673 v4 @ 2.30GHz
<ydalton>
well, from what i found it is old, but still 40 cores
<stintel>
last chromium build was 3h, on -j40
<ydalton>
leio: become a fedora packager at this point
<stintel>
last gcc 1h06
<ydalton>
stintel: wait, last time i built (ungoogled) chromium it was 3h at -j16
<stintel>
leio numbers are way more impressive :)
<stintel>
ydalton: 2.3GHz and old
<ydalton>
true, i'm on a ryzen 5800X i think
<ydalton>
my cousin got a mac studio and i'm jealous :P
<ydalton>
leio: how long does a kernel build take?
<leio>
1m06
<ydalton>
holy shit
<ydalton>
i'd love to see a video of the modules being built so fast
<leio>
I'm unsure how much stuff is in there (compared to a allmodconfig or whatnot), just what I got from chadmed to start off with
<ydalton>
flying
<leio>
gcc got a bunch of build speed improvements done in gentoo like the last 2 days
<leio>
so that's not only the machine if your builds have been older
<stintel>
ah, interesting. my last build was 6 days ago
<leio>
ryzen 3700X is what I have otherwise
<leio>
on slower machines you should see quite a speed increase, especially on slower machines
<leio>
gcc has been quite a bad thing for me, I went clang now like chadmed
<ydalton>
i don't know the difference between either, i just use gcc
dhmltb^ has joined #asahi
<leio>
gcc doesn't even optimize at all for the machine, you get some basic armv8-a only if you specify by hand or specify BOTH -march=native and -mcpu=native
<ydalton>
are there even apple silicon specific optimizations in the compilers yet?
<leio>
the feeling is that llvmpipe is quite a bit faster than it was on the temporary fedora install
<leio>
they are there in llvm/clang, not for gcc
<ydalton>
hm, maybe a reason to consider clang
<leio>
my workaround list is a bit long though
<leio>
this is becoming #asahi-alt territory really though
<ydalton>
what does supertuxkart on an m2 studio with llvmpipe look like?
<leio>
I don't know, videos were very choppy though. I'm getting to daily driving it in practice the upcoming week
<leio>
running 4K probably isn't helpful
<ydalton>
really? i remember even on software rendering videos were okay
<leio>
I'll see soon with clang optimized stuff; I checked it out on previous install I wiped due to bad partioning as I had thought the install will be temporary, but turned out better than expected
<leio>
the previous one was gcc still at the time
<leio>
I think Fedora would be building with some baseline explicit -mcpu/-march, so it was probably worse for me back then - just had found out the gcc bug from a ffmpeg configure failure
<leio>
so my stuff was all very much just -O2 -pipe in gentoo
<leio>
I can't understand these partioning sizes. Tells me 1TB free, I go put it to use for /home, I get 932GB
<ydalton>
haven't you seen that stuff before?
<ydalton>
it's the GiB vs GB difference
<leio>
looks like it, so asahi installer talks in the 1000's
ydalton has quit [Remote host closed the connection]
<leio>
I just managed to wipe my install partially from curl'ing the wrong thing from history, any good way to restore while keeping the untouched actual linux rootfs in place?
<leio>
so ESP and rootfs is still there, but need to re-do the 2.5GB APFS thing :(
<janneg>
hard to say without knowing what you actually did. deleting macos stub and esp and reinstalling uefi only + https://github.com/leifliddy/asahi-fedora-usb should be able to rescue things
<leio>
the shell history was out of order and I ran wipe-linux and quickly cancelled it
<janneg>
if the ESP is still intact, save the ESP contents, delete the sub and esp, do uefi-only install, restore ESP and be prepared to fixing fallout from the changed ESP PARTUUID
<leio>
I'm not sure I can do uefi-only install as that's available in the arch installer and that doesn't work on my device due to fw versions or some such
<j`ey>
leio: put it in your custom installer too?
<leio>
so I should grow back the 1TB I wanted to make my /home on, so that install doesn't put things in the wrong place, and then go from there
<leio>
so I just update the json to have these version numbers and it'll be fine?
<leio>
fw version numbers
<janneg>
or safer. keep the ESP and resize the macos partition so that you have enough space
<janneg>
yes, the u-boot in m1n1/boot.bin will not work, copy the one from your ESP
brunsgaard has joined #asahi
jeffmiw has quit [Ping timeout: 480 seconds]
<leio>
further
<leio>
ok, so; 1) I'll take away the exta space so the new install wouldn't make the new APFS/EPS in the wrong place on the disk; 2) I'll back up ESP just in case (I only have my custom boot.bin that's already backed up anyways); 3) delete APFS and ESP; 4) run a hacked up alarm installer that uses a json that claims 13.5 fw to get the APFS back and not have to separate the ESP stuff; 5) fix up m1n1's re M2 patches and ESP UUID; 6) call it a night before I mess up things
<leio>
what if I made a installer_data.json that doesn't even install any disks at all?
<leio>
partitions*
brunsgaard has quit [Ping timeout: 480 seconds]
<leio>
janneg: also, would it be the same if I just made a version that uses my zip but doesn't specify a root partition in the entries and thus would hopefully just ignore the root.img?
<j`ey>
leio: this is why it's probably better to do most of the actual dev work on another machine, so you can just blast all the disk awya without issue :)
<leio>
I was setting this up to be my actual dev machine :(
<j`ey>
yeah but you should do that after you get it all working!
<leio>
my bad, but might be nice to have a little "type wipe" prompt in there :)
<leio>
oh, it was all working, I even have full GNOME on it, was just making a /home partition in mac OS
ydalton has joined #asahi
<leio>
but then just had to go back to look into GB vs GiB, but the mac terminal was kind enough to give me wipe-linux on Ctrl+R in the history, even though that was ran long before the last installer runs
<ydalton>
leio: fedora asahi creates a /home subvolume in the same / partition
<ydalton>
leio: i've had things like that happen to me before, but nothing nuclear like this
<ydalton>
reading the script itself, you must have pressed enter twice to activate its fury
<leio>
maybe
<ydalton>
although, type wipe would have been a good safeguard
<leio>
just a suggestion, I realize it's all on me there
<leio>
don't do stuff like that at 1am in the middle of IRC convos :)
<leio>
at least the Preboot/Recovery volume setups are going 8+ MB/s again, not 320kB/s like one random time - that was weird